Menu
Back to AI and Diagnosis
02/25/26

Interpretability of an FDA-authorized AI/ML sepsis diagnostic tool improved by SHAP values.

Watson, Gregory L; Staples, Grace; Carver, Robin; Bhargava, Akhil; López-Espina, Carlos; Schmalz, Lee; Ali, Farhan; Antkowiak, Peter S; Azad, Saleem; Berghea, Ramona; Chawla, Lavneet; Crisp, Matthew; Dagan, Alon; Davila, Francisco; Davila, Hugo; DeMarco, Carmen; Doodlesack, Amanda; Espinosa, Aimee; Evans, Neil S; Ezekiel, Clinton; Friederich, Andrew; Gosai, Falgun; Halalau, Alexandra; Iyer, Karthik; Kravitz, Max S; Kurtzman, Niko; Lee, John H; Maddens, Nicholas; Malkani, Roneil; Mayer, Stockton; Oke, Vikram; Palagiri, Ashok V; Patel, Roshni; Raghavakurup, Lekshminarayan; Raouf, Samuel; Reseland, Eric; Sadaka, Farid; Sarma, Deesha; Smith, Scott; Shvilkina, Tatyana; Sims, Matthew D; Singh, Sahib; Stenson, Bryan A; Syed, Anwaruddin; Tafa, Muleta; Thomas, Kurian; Zhao, Sihai Dave; Zhu, Ruoqing; Bashir, Rashid; Reddy, Bobby; Shapiro, Nathan I. JAMIA Open.

Collection Topics
AI and Diagnosis Big Three Conditions
Share

Most recent

Primary care signals and diagnostic pathways before incident heart failure: a systematic scoping review.

Jerjes, El-Osta, Jhass, et al., J Prim Care Community Health.
09/09/26 Big Three Conditions Culture, Work System, and Implementation Epidemiology

Emergency department revisit quality review screening: exploring human decision-making and artificial intelligence support.

Handler, Robles-Granda, Mefford, et al., arXiv
09/09/26 AI and Diagnosis Measurement, Monitoring, and Event Analysis

A five-phase evaluation framework for diagnostic and predictive medical artificial intelligence.

Ye, Chen, Huang, et al., NPJ Digit Med.
09/03/26 AI and Diagnosis

Summary

AI tools in healthcare can enhance patient care, but their lack of transparency makes clinicians cautious. This study examines whether standardized explanations added to an FDA-approved AI sepsis tool improved clinician understanding and trust. Clinicians interpreted explanations accurately 98% of the time, preferred having them, and gained insight into how sepsis risk scores were generated. Standardized explanations can make AI tools more transparent and encourage their adoption in patient care.

Read the paper
Watson GL, Staples G, Carver R, et al. Interpretability of an FDA-authorized AI/ML sepsis diagnostic tool improved by SHAP values. JAMIA Open. 2026;9(1):ooag020.